Hugging Face's logo

lsh9034
/
ci-net

PyTorch
Safetensors
meteorology
convection-initiation
satellite
radar
simvp
nowcasting
Model card Files
xet
 Community
ci-net
File size: 2,751 Bytes
76d61a0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
device: auto
output_dir: ../../../result/training/reference
seed: {value: 42, deterministic: false}
logging:
  wandb: {enabled: false}
dataset: {type: fast}
output_root: ../../../result/final_preprocess/training_full
catalog_path: ../../../result/final_preprocess/training_full/catalog.csv
input_window: {past_minutes: 50, interval_minutes: 10}
input_sources: [concat, hsr]
labels:
  bt: {future_minutes: 60, interval_minutes: 10}
train:
  split: train
  time_ranges:
    - {start: "201907010000", end: "201910312350"}
    - {start: "202005010000", end: "202010312350"}
    - {start: "202105010000", end: "202110312350"}
    - {start: "202205010000", end: "202210312350"}
    - {start: "202305010000", end: "202310312350"}
  input_sources: [concat, hsr]
  required_labels: [ci_hard, bt]
  epochs: 250
  batch_size: 3
  num_workers: 2
  persistent_workers: true
  prefetch_factor: 2
  shuffle: true
  drop_last: true
  grad_clip_norm: null
  use_ram_cache: false
  use_ram_chunk: true
  resume: true
  resume_path: null
valid:
  split: valid
  time_ranges:
    - {start: "202405010000", end: "202410312350"}
  input_sources: [concat, hsr]
  label_key: ci_hard
  required_labels: [ci_hard]
  batch_size: 2
  num_workers: 2
  shuffle: false
  drop_last: false
  use_ram_cache: true
  use_ram_chunk: false
ram_cache: {use_ram_cache: false}
ram_chunk:
  enabled: false
  chunk_ram_gb: 44.0
  async_prefetch: true
  read_block_rows: 64
  cache_dtype: float16
  swap_policy: repeat_current
  chunk_order: circular
  verbose: true
dataloader: {pin_memory: true}
model:
  name: simvp_bt_aux
  params:
    in_shape: [6, 12, 583, 550]
    out_shape: [1, 1, 583, 550]
    bt_out_shape: [6, 1, 583, 550]
    hid_S: 64
    hid_T: 256
    N_S: 6
    N_T: 8
    model_type: gSTA
    groups: 1
    mlp_ratio: 8.0
    drop: 0.1
    drop_path: 0.0
    spatio_kernel_enc: 3
    spatio_kernel_dec: 3
loss:
  name: ci_bt
  params: {normalize_terms: true, ref_batches: 100, eps: 0.00000001}
  ci_loss:
    name: bce
    output_key: ci
    label_key: ci_hard
    weight: 0.85
    params: {eps: 0.000001}
  bt_loss:
    name: bt_mse
    output_key: bt
    label_key: bt
    weight: 0.15
    params: {}
optimizer: {name: adamw, params: {lr: 0.001, weight_decay: 0.0001}}
scheduler: {name: step, params: {step_size: 10, gamma: 0.9}}
validation:
  checkpoint_glob: null
  start_epoch: null
  checkpoint_stride: 2
  max_checkpoints: null
  output_dir: null
  eval_workers: 16
  thresholds: [0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9]
  cluster_mode: mode_like
  min_cluster_pixels: 3
  label_min_cluster_pixels: 0
  merge_buffer_km: 12
  label_buffer_km: 0
  model_false_buffer_km: 0
  cluster_mask_expansion_km: 0
  pixel_size_km: 2.0
  connectivity: 8
  copy_best_model: true